Federated Louvain Community Detection With Secret Sharing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Constructing a user-user graph in a cross-regional environment is challenging due to restricted data sharing between regions, necessitated by privacy requirements, making traditional community detection algorithms infeasible.
Innovation Solution
Implementing a federated Louvain algorithm with secret sharing technology to securely partition and analyze user-user graphs across regions, using distributed computing and secret sharing techniques to protect data privacy while enabling community detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional community detection algorithms are used, then community detection can be performed, but data privacy requirements cannot be satisfied due to unrestricted data sharing
Solution Approach 1:
The patent segments the graph data into multiple partitions distributed across different regions, with each region holding only local data. The Louvain algorithm is then adapted to operate on these segmented partitions, performing community detection locally while preserving data privacy. This segmentation allows community detection to proceed without requiring centralized access to all user data.
Solution Approach 2:
The patent introduces an intermediary mechanism using secret sharing technology and secure multi-party computation protocols. These intermediaries enable collaborative community detection across regions without directly sharing sensitive user data. The intermediaries facilitate information exchange in a privacy-preserving manner, allowing the algorithm to function despite data distribution constraints.
2Measurement precision
If data is shared across regions for community detection, then detection accuracy improves, but privacy requirements are violated
Solution Approach 1:
Secret sharing technology acts as an intermediary that enables accurate community detection while preventing direct access to sensitive user data. The technology allows regions to collaboratively compute community structures without exposing individual user information, thus maintaining both detection accuracy and privacy protection.
Solution Approach 2:
The patent creates copies of graph data in encrypted or secret-shared form across different regions. These copies enable distributed computation of community detection without requiring access to the original sensitive data. The copying mechanism allows accurate detection while preserving privacy through cryptographic protection.
3Productivity
If centralized graph processing is used, then processing speed is high, but data distribution and privacy constraints cannot be met
Solution Approach 1:
The patent segments the graph processing task into distributed operations across multiple regions. Each region processes its local partition independently using the Louvain algorithm, enabling parallel computation that maintains high processing speed while adapting to data distribution constraints. The segmented approach allows simultaneous processing without centralized bottlenecks.
Solution Approach 2:
The patent implements a dynamic federated Louvain algorithm that adapts to varying data distributions across regions. The algorithm dynamically adjusts its operation based on local data characteristics while maintaining overall consistency. This dynamic approach enables the system to handle diverse data distribution scenarios while preserving processing efficiency.
Data Source
AI summary
A computer-implemented method includes accessing, by one of more devices of a first region, an input graph comprising a plurality of nodes and a plurality of edges, each edge connecting two nodes from the plurality of nodes, where each node represents one or more users from the first region. For each node and using a secret sharing protocol: 1) one or more modularity gains for moving the node from an original community into one or more respective candidate communities is calculated and 2) an identified direction for moving the node based on the one or more modularity gains is calculated. The input graph is partitioned into a plurality of communities based on moving each node in the respective identified direction. If a determination is made that a threshold condition has been satisfied, an output graph is generated for the plurality of communities.